0
$\begingroup$

When searching about noninformative priors on internet, one can read here and there that those priors in fact turn out to be informative. However, I did not yet read a real argument about that.

So my question: in what noninformative (typically Jeffreys') priors turn out to be informative ?

Improve this question
$\endgroup$
3
  • $\begingroup$ Check the thread I marked as a duplicate of yours. It has a good discussion of this topic. TL;DR there is no such a thing as "uninformative" prior, priors always carry some sort of information. $\endgroup$
    – Tim ♦
    Nov 8, 2021 at 10:08
  • $\begingroup$ Thanks for the answer. I was aware of that discussion. May I ask here: you say "there is no such a thing as "uninformative" prior, priors always carry some sort of information" But how do you define what does "carry information" mean for a prior ? Because without explanation that seems a bit dogmatic to me (and it was precisely the topic of my question) Maybe I can open a new discussion with this question ? $\endgroup$
    – Celi
    Nov 8, 2021 at 11:28
  • $\begingroup$ this is answered in the linked thread. But simple answer: uninformative prior should not influence the result and that is not the case, since the choice of prior always has some effect on the result. $\endgroup$
    – Tim ♦
    Nov 8, 2021 at 11:54

0

Reset to default

Browse other questions tagged or ask your own question.